GitLab

* add keep_training_booster

* use model string

* reset handle; free dataset

thanks for building in early_stopping! super useful.

* setup.py custom install & sdist

* global data_files

* fix install rule

* fix pylint

* clean codes

* use distutils.command

* overwrite install_lib

* add copy_files; fix pylint

* install openmpi only TASK=mpi

* add TASK=sdist & bdist

* setuptools.command instead of distutils

* add METHOD=pip for gpu test

* use openmpi to install gcc

* set OpenCL_INCLUDE_DIR & BOOST_ROOT

* update doc

* export BOOST_ROOT

* update doc

* update doc

* update doc

eval_at has been broken after e465f92a.
This commit will fix eval_at.

* add make command to the python package.

* Update README.rst

* Update README.rst

* Update README.rst

* fix tests.

* fix unix build

* update readme

* fix setup.py

* update travis

* Update .travis.yml

* Update test.py

* some fixes.

* check the 64-bit python

* fix build.

* refine MANIFEST.in

* update Manifest.in

* add more build options.

* Add fatal in cmake

* fix a endif.

* fix bugs.

* fix pep8

* add test for the pip package build

* add test pip install in travis.

* fix version with pre-compile dll

* fix readme.rst

* update readme

* handle list for prediction

* fix error msg

This commit will regain backward compatibility that has been broken
after e465f92a.

* check parameters

* refine sklearn verbose logic

* fix segfault

* Warning -> Fatal

* fix segfault

* support kwargs for sklearn

* add link to Parameters

* fix multi-threading.

* fix name style.

* support in CLI version.

* remove warnings.

* Not default parameters.

* fix if...else... .

* fix bug.

* fix warning.

* refine c_api.

* fix R-package.

* fix R's warning.

* fix tests.

* fix pep8 .

* Add early stopping for prediction

* Fix GBDT if-else prediction with early stopping

* Small C++ embelishments to early stopping API and functions

* Fix early stopping efficiency issue by creating a singleton for no early stopping

* Python improvements to early stopping API

* Add assertion check for binary and multiclass prediction score length

* Update vcxproj and vcxproj.filters with new early stopping files

* Remove inline from PredictRaw(), the linker was not able to find it otherwise

This commit follows Microsoft/LightGBM@3abff37

* make test fail

* change default best_iteration to 0

* fix test

* change data_splitter to folds in cv

* update docs

* python-package: fix creating eval_set in LGBMClassifier

* replace elements in eval_set directly

* change whitespace to underline in feature names

* add test

* fix bug

* fix bug

* warning -> fatal